Claude AI for Marketing
Learn how to use the Claude to level up your day to day digital marketing and growth, build reusable marketing workflows and systems with Claude Skills, MCP, Claude Code or Claude Cowork, as well as…
- 53.2K likes on YouTube
- 2.1M views
Created by Grace Leung on YouTube
What you'll learn
- Claude Code • Building AI agents
- Claude Design • Brand design systems
- Claude Cowork • Setup and workspace
- Marketing • Skill types and build order
- Claude Code • Project setup and folder structure
- Marketing • Content system architecture
- Claude Skills • Building a skill from SOPs and references
- Claude • Desktop work modes (chat, Cowork, Claude Code)
Explore related topics
Course content
01Claude Code: Build Your First AI Agent in 20 Minutes (100% No Coding)Preview20:01
This chapter shows how to build AI agents with Claude Code, a desktop app that does not require programming. An agent is defined by its ability to make decisions toward a goal: which steps to take first, which tools to use, and whether the result is good enough. Every agent needs five parts: a brain (the AI model), context (what it knows about your brand), a task (its goal), rules (how it works and what it must never do), and tools (what it is allowed to use). The workspace is simply a folder that holds brand files and reference posts, and Claude Code's permission mode controls how often it asks for approval. The instructor builds a content agent that turns long content into carousel posts. The agent's quality improves through three steps: adding a training guide and brand context, creating skills (step-by-step playbooks for repeatable output), and adding memory so it learns from feedback over time.
- Claude Code • Building AI agents
- Agents • The five parts of an agent
- Claude Code • Workspace and permissions
- Claude Code • Skills and memory
- Claude Code • Cloud agents and scheduled tasks
02Master Claude Design in 17 Minutes for MarketingPreview16:43
This chapter explains how marketers can use Claude Design, Anthropic's visual design studio, as a reusable system rather than starting from scratch each time. The core idea is to build a brand design system once (logo, colors, fonts, and one real example asset), then generate templates from it, and finally package those templates as reusable Claude skills that can run anywhere inside Claude. The workflow is to keep brand context and instructions in Claude projects, use Claude Design mainly for visual decisions, and let execution happen in Claude.
- Claude Design • Brand design systems
- Claude Design • Reusable templates and skills
- Marketing • Carousel and presentation templates
- Marketing • Lead magnets and landing pages
- Claude • Packaging design outputs as skills
- Claude • Motion graphics workflow
03Claude Cowork: 8 Use Cases Every Marketer Should Know (Full Tutorial)Preview20:51
This chapter explains how to use Claude Cowork, a local desktop agent, for marketing work. Unlike Claude Chat, which runs in the cloud and needs manual copying and saving, Cowork reads and writes files in folders on your computer and can connect to outside tools. The chapter starts with workspace setup: create a marketing folder with context, template, and project subfolders, then set simple global instructions. Claude can build the brand context, brand voice, and ideal customer profile files by asking you questions, and those files are then reused in every task so you do not repeat yourself.
- Claude Cowork • Setup and workspace
- Claude Cowork • Projects and project instructions
- Claude Cowork • Connectors and live data
- Claude Cowork • Live artifacts
- Claude Cowork • Skills, plugins and browser use
- Claude Cowork • Scheduled tasks and automation
04Claude Design + Claude Skills: Automate Your Marketing (Claude Code)Preview16:49
This chapter shows marketers how to turn Claude into a repeatable marketing system using Claude Skills, Claude Code, and the Claude design tool. The presenter splits marketing work into three skill types: brand skills (brand standards such as voice and visual design), function skills (day-to-day marketing tasks), and specialty skills (rules for a specific domain). The advice is to list the repeated tasks in your week, sort them into these types, and build brand skills first because almost every task depends on them.
- Marketing • Skill types and build order
- Claude Code • Project setup and CLAUDE.md
- Claude design tool • Design system extraction and export
- Marketing • Campaign workflow automation with skills
- Claude • Campaign manager agent
- Marketing • Team skill library and scheduled sync
05Claude Code: Build Your Full AI Marketing Team (Agents + Claude Skills)Preview17:35
This chapter shows how to build a marketing team inside Claude Code, using a travel brand called Go Travel as the example. The process has four steps: map the marketing tasks you repeat every week, turn each repeatable workflow into a skill, group similar skills into focused agent roles, and connect agents and skills as a team through a project setup file. The project is organized into system folders (such as context, SOPs, and templates) that agents reuse, and working folders where outputs go. Setup starts with loading brand context, creating a CLAUDE.md file with project instructions, and installing Anthropic's official skills pack in VS Code.
- Claude Code • Project setup and folder structure
- Claude Code • Building skills
- Claude Code • Agents and routing rules
- Claude Code • MCP tools and image generation
- Claude Code • Task boards and remote control
06Claude Cowork + Skills is Your NEW Content SystemPreview17:16
This chapter shows how a marketer can build a content system with Claude Cowork and Claude Code that works from one project folder. The system has three core layers: a context layer (brand knowledge, voice, audience, products), a skill layer (reusable instructions for specific tasks), and an orchestration layer (Claude connected to tools that reads inputs, runs the right skills, and pushes outputs). The whole system can be driven by a simple spreadsheet, called the content planner, which tracks each content type, its inputs, and its status. The demo brand is Goldie Coffee, and the project folder holds a system folder, a script folder, a working folder, a Claude.md system prompt, and the content planner. Because Cowork does not read project-level skills automatically, the skill folder location must be added to Claude.md or to Cowork global instructions.
- Marketing • Content system architecture
- Claude Cowork • Setup and project instructions
- Marketing • Content repurposing
- Marketing • Carousel and ad creative generation
- Claude Code • Publishing to Notion and WordPress
- Marketing • Blog drafting for SEO
07Claude Skills: Build Your First AI Marketing Team in 16 Minutes (Claude Code)Preview16:43
This chapter shows how to build an AI marketing team using Claude Skills in Claude Code Desktop. A skill is a packaged set of expertise (SOPs, frameworks, brand standards, reference files) that Claude follows whenever it runs that skill. Agents are the team members who use these skills. The presenter argues that you build a skill library once and Claude then works like a team that runs on it. The demo uses a fictional D2C home-goods brand called Simple Home, with a project folder that holds context files and a Claude.md file to guide Claude.
- Claude Skills • Building a skill from SOPs and references
- Claude Skills • Creating skills from examples and frameworks
- Claude Skills • Multi-skill orchestration
- Claude Agents • Subagents vs. Agent Teams
- Claude Plugins • Packaging and reusing skill libraries
- Marketing • Brand context and project setup for AI
08Claude Cowork Just Changed How You Do MarketingPreview17:17
This chapter explains how to use Claude as a working teammate for marketing tasks on your computer. It starts by separating Claude's three desktop modes. Claude chat is for brainstorming and planning, with data stored in Anthropic's cloud. Cowork is described as Claude Code for non-technical users: it runs the same agent engine, can access local files, runs parallel agents, and is built to finish tasks end to end. Claude Code Desktop offers about 95% of Claude Code's power without a terminal, and the Claude Code terminal gives the most control. The recommended approach is to use chat for thinking, Cowork for getting work done, and Claude Code for building, and to combine them when it makes sense.
- Claude • Desktop work modes (chat, Cowork, Claude Code)
- Claude • Cowork end-to-end task execution
- Claude • Parallel sub-agents and batch workflows
- Claude • Custom skills and Claude in Chrome
- Claude • Plugins, commands, and skills
09Claude Code just Built me an AI Agent Team (Claude Code + Skills + MCP)Preview17:23
This chapter shows how to use Claude Code to build a team of AI agents that work together on real tasks, not just chat. The demo uses a fictional marketing agency. The creator starts by making a project folder with business context, then has Claude read that context and create a Claude.md file, which acts as the project's system prompt. Claude Code can be accessed through the local terminal, an IDE extension, the web app, or the desktop app. The terminal is recommended because it gives the most complete set of commands and features. Agents are created with the /agents command, which saves each one as a markdown file in an agents folder. Each file holds a system prompt that defines the agent's role, knowledge, and tools. Agents can also be given templates and standard rules so their output stays consistent.
- Claude Code • Setup and installation options
- Claude Code • Claude.md and project context
- Claude Code • Creating agents with the /agents command
- Claude Code • Skills (official and custom)
- Claude Code • MCP servers and live data
- Claude Code • Multi-agent routing and orchestration
10Master Claude Skills in 18 Minutes | Claude Skills ExplainedPreview18:15
This chapter explains Claude skills, also called agent skills: reusable instruction manuals that tell Claude how a task should be done, which tools to use, and what standard to follow. Before skills, users pasted the same instructions into every chat, or stored them in Claude projects. Projects have two limits: their instructions are not portable outside the project, and projects cannot be combined (not stackable). Skills fix both by being stored at the account level, where they can be used across the web app, Claude Code, and the API, and combined together.
- Claude • Skills basics
- Claude • Skills vs projects vs MCP
- Claude • Types of skills
- Claude • Creating custom skills
- Claude • Using skills in workflows
- Claude • When to create a skill
11Unlock the 90% of Claude + Notion Power Most People OverlookPreview16:03
This chapter shows how Claude can be combined with Notion to make a Notion workspace smarter and more useful for marketers and working professionals. Setup is simple: in Claude's web app or desktop app, the user adds the Notion MCP connector from the directory, connects it, and selects the Notion workspace to share. Once connected, Claude can read and write Notion content directly from the chat.
- AI Tools • Claude + Notion MCP setup
- Marketing • Content calendar automation
- Marketing • Design with Canva MCP
- Project Management • Dashboards from Notion data
- Operations • Team knowledge base and onboarding
- AI Tools • Data enrichment and limitations
12Why MCP Saves 90% of Your Manual AI Work | Model Context Protocol ExplainedPreview14:57
This chapter explains the Model Context Protocol (MCP), an open source standard that defines how applications give context and tools to large language models (LLMs). Without MCP, connecting an AI to each business tool, such as Salesforce or Google Analytics, requires a custom integration, which is slow to build and does not scale when models or tools change. MCP acts as a universal translator: tools and services that support MCP speak one common language, so an AI only has to learn one protocol. Because of this, AI applications become model agnostic, meaning a team can switch between models without rebuilding integrations from scratch.
- AI • What MCP is and why it matters
- AI • MCP architecture: servers and clients
- AI • Choosing between the four connection methods
- AI • Installing an official MCP server (Notion demo)
- AI • Building a custom MCP server with n8n
- AI • Best practices: permissions and tool limits
13Claude 4 × Gemini 2.5: The Workflow Most People MissPreview18:23
This chapter argues that no single AI model is the best at everything, so skilled users should combine models. The speaker sees Gemini 2.5 as strong at scale and speed, thanks to its very large context window of one to two million tokens, and at fast, technical, data-heavy coding for prototypes. Claude 4 is presented as strong at strategy, storytelling, careful attention to detail, precise analysis, and psychologically aware, polished output. The recommended pattern is to let Gemini do the heavy lifting, such as processing large datasets, analyzing multimodal inputs, and producing a first-pass analysis, and then let Claude turn that analysis into precise insights, compelling narratives, and visually strong output.
- AI Tools • Gemini 2.5 vs Claude 4 strengths
- AI Workflows • Multi-model orchestration
- AI Workflows • Using Gemini for analysis and Claude for strategy
- Data Analysis • Strategic dashboards from large datasets
- Marketing • Deep audience intelligence and personas
- Communication • AI-assisted speaking coaching
14Turn Claude to Powerful AI Agents, Automate 50% of Your Work (Claude + MCP Agents)Preview17:17
This chapter explains how to turn Claude from a chat assistant into an AI agent that can take actions, using MCP (Model Context Protocol). MCP is described as a universal plug: a standard way for AI models like Claude to connect to outside data sources and tools through an MCP server. Without it, Claude could only see data a user uploaded into the chat. With it, Claude can retrieve far more context and perform real tasks. The presenter outlines four ways to get MCP servers: official native integrations (such as Google Drive, Calendar, and Gmail), official MCP servers from vendors (such as Notion and Perplexity), community-built servers (such as Figma), and self-built servers made with no-code tools like n8n. Claude Desktop is recommended for the best MCP support.
- AI • What MCP is and how it works
- AI • Ways to connect Claude via MCP
- AI • Framework for building Claude agents
- AI • Research and business intelligence agents
- AI • Personal assistant and UX design agents
- AI • Project agents and human oversight
15Perplexity + Claude 3.7 will Transform Your Work to NEXT LEVELPreview16:05
This chapter shows how to combine Perplexity and Claude 3.7 Sonnet into one workflow for professional work. Perplexity is used first to gather real-time sources, data points, and reports, using focus modes, academic or social search, or Deep Research. Claude is used second to analyze those sources, think through them, and produce deliverables. Claude cannot browse the internet and has a knowledge cutoff, so Perplexity fills that gap. Claude's Projects keep shared context and instructions, and its Artifacts create interactive outputs such as dashboards, pitches, and mockups.
- AI Tools • Perplexity and Claude workflow
- Claude • Extended thinking mode
- Claude • Projects and Artifacts
- AI Workflow • Visual dashboards from research
- AI Workflow • Research report automation
- AI Workflow • Audience style guidelines and pitches
Description
Learn how to use the Claude to level up your day to day digital marketing and growth, build reusable marketing workflows and systems with Claude Skills, MCP, Claude Code or Claude Cowork, as well as other Claude's ecosystem building blocks.
About the creator
- YouTube channel
- 151K subscribers
- 1 course on Studyio
The videos in this course are made by Grace Leung and play from YouTube, credited to them. Studyio adds the chapters, notes, quizzes and a certificate around them.
More courses like this

Barbara O'Neill Part 1. DNA and The True Cause of Disease

Claude – Full Course 2026 (How to use Claude AI for beginners)

How to Speak

Gilbert Strang lectures on Linear Algebra (MIT)

ROYAL ENFIELD HIMALYAN 450 INSTRUMENT CLUSTER FULL DETAIL ALL SETTING PHONE CONNECTION TRIPPER DASH
